|
BSOD on new X99X install |
Post Reply 
|
Page <123> |
| Author | |
Franko 
Newbie 
Joined: 19 Dec 2015 Status: Offline Points: 18 |
 Post Options
 Thanks(0)
 Quote Reply
 Posted: 06 Jan 2016 at 9:31am |
|
I looked up question on GPU temps.
The Nvidia site has specs on all their gear. The link: http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-780/specifications So I think I am ok, although I can check again after the memory test. Looks like 4 hours per Pass, btw, and I am only through the first pass (with no errors yet) Here is a copy of the GTX 780 specs: Thermal and Power Specs: 95 CMaximum GPU Temperature (in C) 250 WGraphics Card Power (W) 600 WMinimum Recommended System Power |
|
 |
|
|
Xaltar
Moderator Group 
Joined: 16 May 2015 Location: Europe Status: Online Points: 26903 |
Post Options
Thanks(0)
Quote Reply
Posted: 06 Jan 2016 at 7:18am |
|
No, bottlenecking will not cause the GPU to alter it's clocks in any way. The bottleneck occurs when the GPU is starved for data by an inadequate CPU or PCIe bandwidth, this results in stutter and spike behavior. Until fairly recently reviews did not even test for this. Quite often you will see a higher average FPS and a stupidly high maximum FPS when a bottleneck is present but the minimum FPS will be very low and occur frequently. The result, despite the highish average will be very choppy gameplay with loads of stutter. The high FPS spikes will occur immediately after a stall as the CPU suddenly dumps a load of frame data onto the GPU which up till that moment was idling waiting for data. I suppose depending on how long the GPU was starved for data it may briefly lower it's clocks a performance level but it would be for no more than a second or so, not enough to significantly affect thermals.
There is also no way that CPU would bottleneck a GTX 780 and no way the GPU should be running hotter on the new system unless there is some kind of issue.
Is the BSOD the only error you are getting Franco? Do you get any "the display driver has stopped responding and been recovered" messages? This is usually the message you would get if the GPU is broken, defective or the driver is bad. IRQ not less or equal messages are usually memory or hard disk related so it may be worth while reseating your CPU an checking for bent/misaligned pins while you are at it. I have seen quite a few issues on X99 platforms that were resolved by reseating the CPU. Make sure your CPU cooler is not installed so tightly that it is causing the pins to miss-align. Good luck |
|
|
|
|
wardog
Moderator Group
Joined: 15 Jul 2015 Status: Offline Points: 6447 |
Post Options
Thanks(0)
Quote Reply
Posted: 06 Jan 2016 at 6:35am |
|
And that review, while dated now, was with an i7-3960X on a superb EVGA X79 SLI motherboard. No slouching there, that system didn't hamper the card in anyway.
I am not a GPU expert either, your old motherboard and CPU was "older', without knowing what you had previously. I don't understand bottlenecking, yet I can not believe for one second the current system it's in would be a bottleneck to it. Anyone???? I haven't a clue. Does bottlenecking cause the card to physically slow itself in recognition of it being hampered by a/the lowered bus throughput? On whatever the OPs prior system was. |
|
|
|
|
wardog
Moderator Group
Joined: 15 Jul 2015 Status: Offline Points: 6447 |
Post Options
Thanks(0)
Quote Reply
Posted: 06 Jan 2016 at 6:21am |
|
A quick Google search and this is the best I've found concerning GTX 780 temps.
http://www.anandtech.com/show/6973/nvidia-geforce-gtx-780-review/19 Interesting snippet from the above page:
Edited by wardog - 06 Jan 2016 at 6:22am |
|
|
|
|
Franko
Newbie
Joined: 19 Dec 2015 Status: Offline Points: 18 |
Post Options
Thanks(0)
Quote Reply
Posted: 06 Jan 2016 at 5:49am |
|
I appreciate all input. The gear is all new - no dust, except a bit on the reused video card, which was blown off.
I see the GPU temps do go up when I run WOW. I installed one of those apps that show it and the CPU temp continuously on a little bar graph, which I run on a second monitor. I am trying to remember the number I reach on the GPU - maybe 75C, or maybe 80C? Whatever it was, as soon as I started playing the game it went right up to that temp and then was a level line. I think the card prevents it from getting any higher by limiting the work it does, but I am not a GPU expert. The card worked in my Windows 7 machine for over 6 months, but this has a faster CPU so are you thinking that maybe it makes the card run faster/hotter? I am running a full 4 pass memtest, so I can't check right now. Could a hot video card cause a 'IRQL NOT LESS OR EQUAL' BSOD do you think? Thanks, Frank |
|
|
|
|
wardog
Moderator Group
Joined: 15 Jul 2015 Status: Offline Points: 6447 |
Post Options
Thanks(0)
Quote Reply
Posted: 06 Jan 2016 at 5:15am |
|
Too, in just now going back over the thread, check what your GPU temps are. This will also show its ugly head if the GPUs thermal paste has dried out some and now letting the card get warm/hot.
Reading over your posts with your writing style I assume the machine has been cleaned and now free of dust. But I'll ask anyways, no harm no foul in asking, right? |
|
|
|
|
Franko
Newbie
Joined: 19 Dec 2015 Status: Offline Points: 18 |
Post Options
Thanks(0)
Quote Reply
Posted: 06 Jan 2016 at 4:09am |
|
Update:
I removed all the devices, even the CDROM, except for the SSD, the Video card, keyboard and mouse. I removed the M.2 PCIE Samsung SSD and installed only a single 500G SATA SSD. Reseated the memory sticks. I used diskpart to clean the SSD and installed a fresh Win 10 Pro on it in UEFI mode. Install went well, so I updated the drivers, triple checking only those from the ASROCK website. I updated the drivers for Nvidia card with the recommended dec 2015 drivers. I installed minimal apps - Chrome and World of Warcraft (great test case) Played WOW last night for about 3 hours with no issues. Booted machine today, did a couple of web browses and got a BSOD. Problem is stil there. I will do an extended memory test but am beginning to think I have an unstable motherboard. The Nvidia card has been working in another machine for over 6 months with no problems. I will update after memtest completes 4 full passes. thanks guys! Frank |
|
|
|
|
Franko
Newbie
Joined: 19 Dec 2015 Status: Offline Points: 18 |
Post Options
Thanks(0)
Quote Reply
Posted: 06 Jan 2016 at 4:00am |
|
Did you find any errors after 40 hours?
My BSODs usually show up within 4-5 hours of running, sometimes within a few minutes of startup. So I thought 4 or 5 hours of continuous memory tests with no errors would clear the memory. But it only takes time so I will let it run all the way next time. I know I made it to the 4th pass but I don't think it completed all the way.
|
|
|
|
|
PetrolHead
Groupie 
Joined: 07 Oct 2015 Status: Offline Points: 403 |
Post Options
Thanks(0)
Quote Reply
Posted: 05 Jan 2016 at 12:29am |
|
I believe the standard test will do four full passes and I think that's what you should go for, especially since you already got one error. It may take ages, though. With my two sticks and 16 GB of RAM it took nearly 40 hours to finish all four passes. It will also look like it's stuck, as it may not show any progress for 6 hours or more, but as long as the symbols for the CPU cores keep spinning, it's still working through the tests.
|
|
|
Ryzen 5 1500X, ASRock AB350M Pro4, 2x8 GB G.Skill Trident Z 3466CL16, Sapphire Pulse RX Vega56 8G HBM2, Corsair RM550x, Samsung 960 EVO SSD (NVMe) 250GB, Samsung 850 EVO SSD 500 GB, Windows 10 64-bit
|
|
|
|
|
Franko
Newbie
Joined: 19 Dec 2015 Status: Offline Points: 18 |
Post Options
Thanks(0)
Quote Reply
Posted: 04 Jan 2016 at 10:54pm |
|
Brief update:
- Upgraded to 2.1 BIOS. Got the same BSOD later that day. - Downloaded Prime95 and ran it for 20 minutes of so to load CPU 100%. System ran nice and cool and no errors or problems. - Downloaded PassMark MemTest86, installed to USB stick, and ran default. Got a single error after about 15 minutes of running. Stopped the test, checked options, which made me realize I better just use the default test regimen. Restarted for a longer test to see if I could determine location etc of errors. Over the next few days, I ran several hours of mem tests - trying new cold boots, etc, but could not get a single error to reappear. Question for the group: I don't like seeing even ONE memory error, but how many hours of running memtest do you need to decide if that one error was a cosmic ray hit or some other fluke? Future tests: - removed data SSD, backup SSD, all USB devices except Mouse and Keyboard. - continued testing bare machine to see if any BSODs appear - if I get a BSOD, I may remove the M.2 pciE drive and boot to a Sata SSD to eliminate it While I wait: - Evaluate every hardware driver and make sure it is current with no documented issues
|
|
|
|
|
Post Reply
|
Page <123> |
Tweet
|
| Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |
Topic Options
With GPU Boost 2.0, load temperatures are kept tightly in check when
gaming. The GTX 780?�s default throttle